Treatment of Domestic Sewage by Biological Contact Oxidation of Different Packings Under Hypoxia

Desalination and Water Treatment 181 (2020) 188–196 www.deswater.com March

doi: 10.5004/dwt.2020.25036

Treatment of domestic sewage by biological contact oxidation of different packings under hypoxia

Chunhong Shia,b,*, Yao Zhanga,b, Xiaolong Huanga,b, Xinfei Zhanga,b aSchool of Energy and Environmental Engineering, University of Science and Technology Beijing, Beijing 100083, China, emails: [email protected] (C.H. Shi), [email protected] (Y. Zhang), [email protected] (X.L. Huang), [email protected] (X.F. Zhang) bBeijing Key Laboratory of Resource-Oriented Treatment of Industrial Pollutants, University of Science and Technology Beijing, Beijing 100083, China

Received 26 May 2019; Accepted 12 October 2019

abstract Treating low C/N domestic sewage is one of the difficult tasks in wastewater treatment. This study investigated the treatment of low C/N domestic sewage under low dissolved oxygen condition via self-built biological contact oxidation reactor with composite packing and polyurethane suspension packing. Meanwhile, MiSeq high-throughput sequencing was performed on the microbial communities on the surface of the packing to analyze the changes in microbial community structure. The results showed that polyurethane suspension packing reactor had better removal efficacy on + NH4–N, TN and COD than the combined packing reactor, and only the effluent indexes of polyurethane suspension packing reactor could reach the first-class A emission standard of Discharge standard of pollutants for municipal wastewater treatment plant (GB 18918-2002). The removal of biological phosphorus of the single-contact oxidation process was not ideal and could be combined with anaerobic reactor. The richness and diversity of the bacterial community in polyurethane suspension packing reactor was higher than those of composite packing reactor. Polyurethane suspension packing reactor contained relatively richer nitrification-related flora Nitrospira, which provide better growth and reproduction conditions for the denitrification-related flora Betaproteobacteria and Denitratisoma. The large amount of facultative heterotrophic bacteria in Proteobacteria had a significant effect on the removal of organic pollutants. This study provides a theoretical basis for subsequent in-depth research and practical application. Keywords: Low dissolved oxygen; Low C/N domestic sewage; Packing; Biological contact oxidation; Microbial community structure

1. Introduction more carbon sources, which increases the cost of the wastewater treatment [5]. The biological contact oxidation method Water pollution is a primary problem that plagues social is an efficient sewage treatment process, which combines development and survival [1–3]. It is difficult to treat urban the characteristics of activated sludge method and biofilm sewage with low concentration and low C/N ratio in south- method, and has the advantages of simple management, less ern China. There are many problems with the traditional residual sludge, impact load resistance and stable treatment biological method used to treat low C/N domestic sewage, effect. It is widely used in the treatment of industrial waste- such as excessive aeration, insufficient organic carbon source water and decentralized domestic sewage in China [6,7]. Yin and low nitrogen removal efficiency [4]. If aeration is used et al. [8] used anaerobic baffled reactor and two-step biolog- to increase the dissolved oxygen (DO), it is necessary to add ical contact oxidation to treat dyeing wastewater. The COD,

* Corresponding author.

BOD and color of the effluent met the requirements of indus- of the reactor is a plexiglass cylinder of H500 mm◊Φ150 mm trial emission standards [8]. (i.d.) with an effective volume of 8 L, and can be filled Packings can be used to remove contaminants in a vari- with contact packing. One side of the reactor has a water ety of ways, such as adsorption [9], biodegradation, chemical inlet near the bottom, and the other side has a water out- reactions, etc. El-Khateeb et al. [10] used non-woven sheets of let or sampling ports every 90 mm from top to bottom, five polyethylene terephthalate (that could be made from waste in total. The bottom is equipped with a drainpipe. A high- plastic bottles) as packings for UASB/DHNW combined sys- temperature-sintered bubble stone aerator is installed under tem to treat domestic sewage, effectively removing COD, the inlet nozzle near the bottom of the reactor, which can BOD, TSS and fecal coliforms [10]. The choice of packings be connected to an external air pump to realize oxygenation in biological contact oxidation processes is critical, requiring and aeration function. packings with high mechanical strength, hydraulic stability Two reactors was pre-packed and labeled as No. 1 and and specific surface area to allow more biofilms to adhere to No. 2. No. 1 reactor was filled with the composite packing, the packing. Especially under the condition of DO less than and No. 2 reactor was added with polyurethane suspended 2 mg/L, it is still able to attach, grow and maintain the high packing (recorded as 1# packing, 2# packing, respectively). activity of microorganisms [11,12]. The fill ratio of reactors No. 1 and No. 2 was 50% (packing In recent years, molecular biology methods such as gra- volume/reactor effective volume). Both packings (Fig. 2) dient gel electrophoresis, clone library and high-throughput were purchased from a packing plant in Beijing, and their sequencing have been widely used to study the microbial physical and chemical properties are shown in Table 1. Fig. 3 community structure and diversity in wastewater treatment. is the SEM for experimental selected packings. Polyurethane MiSeq high-throughput sequencing is based on Illumina’s suspension packing has a mesh structure inside with many sequencing technology. Simultaneous large-scale parallel voids and a smooth surface. The interior of composite pack- sequencing of millions of gene fragments by reversible termi- ing is branched and dense. nation reagent method enables accurate and rapid analysis of results and is therefore widely used in microbial commu- 2.2. Experimental operating parameters nity structure and diversity in wastewater treatment [13–15]. In this paper, a biological contact oxidation reactor experi- In the experiment, the inlet water of the reactor was mental device was designed. Under the condition of low DO manually distributed. The simulated domestic sewage was

(1.5–2.0 mg/L), the treatment effect of two different packing mixed with NH4Cl, NaNO2, K2HPO4, potassium hydrogen biological contact oxidation processes on low C/N domestic phthalate and glucose. The inoculated sludge was taken sewage was investigated. MiSeq high-throughput sequenc- from the Gaobeidian sewage treatment plant. After the ing was used to analyze the microbial community structure preparation work was completed, the peristaltic pump and in different packings and to provide basic data and theoret- the outlet valve of the tower top were opened, and the sim- ical guidance for the development of a low C/N domestic ulated domestic sewage was taken out from the inlet tank wastewater process with good denitrification effect and low and pumped into the reactor to realize the continuous water energy consumption. inlet and outlet. The air pump was turned on, and the DO content in the reactor was controlled by the rotor regulating flow meter. 2. Materials and methods The biofilm was cultured from activated sludge inoc- ulation in the start-up phase of the bioreactor. We first 2.1. Experimental device and process used a high-concentration sewage to grow the biofilm, The experimental device and process of the biological and then used a low-concentration sewage to acclimate contact oxidation reactor are shown in Fig. 1. The main body the biofilm. After the reactor was operated for 30 d, the

Outlet 1

Filler 2 Sampling port layer Oxidation reactor 3

4 Inlet Intake tank 5 Flowmeter Peristaltic pump Aeration pump Draining mud

Fig. 1. Experimental device and process of the biological contact oxidation reactor. 190 C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196

(a) (b)

Fig. 2. Experimental selected packings ((a) composite packing and (b) polyurethane suspension packing).

Table 1 Physical and chemical properties of packings

Sort Size (mm) Specific surface Porosity area (m2·m–3) (%) Composite packing Φ150 290–310 96 Polyurethane suspension packing D20 1,000–1,200 >99

a b

Fig. 3. SEM for experimental selected packings ((a) composite packing and (b) polyurethane suspension packing). adhesion of the biofilm was enhanced, and the biochemical function became more efficient. At this time, the effluent 2.3. Analysis methods water became gradually clearer. When the removal rate + 2.3.1. Determination and analysis of water quality indicators of COD and NH4–N in the effluent was greater than 60%, the biofilm cultivation is successful [16]. After the reactor The concentration of COD in the influent and efflu- was stably operated, biological samples of the two packing ent water was determined by potassium dichromate rapid + surface biofilms were collected and analyzed for microbial digestion spectrophotometry. The concentration of NH4–N composition. The operating parameters and influent water was determined by Nessler’s reagent spectrophotometry. quality indicators of the biological contact oxidation reac- The concentration of TN was determined by alkaline potas- tor are shown in Tables 2 and 3. sium persulfate digestion ultraviolet spectrophotometry. C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196 191

Table 2 40 100 Process operating parameters 35 HRT (h) Influent flow (L/h) DO (mg/L) Temperature (°C) 80 30

6 1.34 1.5–2.0 24 ± 2 ) 25 60 Table 3 20 InfluentInfluent watewater -N(mg/L) Influent water quality indexes -N(mg/L) 1#1# effluenteffluent watewaterr + + 4 4 2#2# effluenteffluent watewaterr 40 15 1#1# reremmoovalval rate NH NH Removal rate(% Indexes Range Average 1#1# reremmoovalval rate Removal rate(%) pH 7.08–7.91 7.47 10 20 + NH4–N/(mg/L) 10.52–56.74 27.35 5 TN/(mg/L) 17.01–72.51 32.16 COD/(mg/L) 66.41–225.87 124.26 0 0 0102030405060 TP/(mg/L) 3.24–6.15 4.89 Day

+ The concentration of TP was determined by potassium per- Fig. 4. Effect of two kinds of packings on the removal of NH4–N sulfate digestion molybdenum antimony anti-spectropho- from wastewater. tometry. The absorbance was measured using HACH DR5000 UV-visible spectrophotometer [17], and the DO concentra- greatly, but Fig. 4 shows a trend of decreasing first, then tion was measured with a HI9146 portable DO meter (the increasing and finally stabilizing. In general, 2# packing measurement position was in the middle of the packed bed). + was better than 1# packing for NH4–N removal. In the early + stage of the reaction, the removal rate of NH4–N of 1# pack- 2.3.2. Determination and analysis of microbial composition ing and 2# packing was low because the biofilm had not yet Two different packing surface biofilms were taken as fully formed, and the growth cycle of nitrifying bacteria was samples for biodiversity measurement. After sampling, the slower, and the adaptation period was long. Therefore, nitri- samples were stored in a –80°C refrigerator and tested in time. fying bacteria were at an inferior stage in competition with DNA extraction and PCR amplification: The DNA was heterotrophic bacteria. extracted using the procedure provided in the EZNA Soil During the whole experiment, the lowest removal rate DNA Kit (Omega, USA) instructions, and the extracted of 1# packing was 51.52%; the highest removal rate was DNA was detected by 1% agarose gel electrophoresis [18,19]. 78.11% and the final removal rate was stabilized at 77.61%. The amplified region of PCR was the V3-V4 region of 16S The lowest removal rate of 2# packing was 51.21%; and rRNA, and the primer for bacterial 16S rRNA amplification the highest removal rate was 90.11% and the final removal was universal primer (338F/806R). The PCR amplification rate was stabilized at around 89.22%. The reason may be system contained 30 ng DNA sample, 1 μL forward primer that 2# packing has a special mesh structure and a higher (5 uM), 1 μL reverse primer (5 uM), 3 μL BSA (2 ng/μL), void ratio than that of 1# packing, which is more condu- cive to the retention of NH+–N. As the reaction proceeds, 12.5 μL 2xTaq PCR reagent, and 7.5-X μL sterile double dis- 4 tilled water. Reaction procedure: the system was preheated the pores of the packing began to become smaller due to at 95°C for 5 min, and then carried out 28 cycles (95°C defor- the adsorption and clogging of the sludge and microorgan- mation 30 s, 56°C annealing 30 s, 72°C extension 40 s), and isms, and the removal rate began to decrease. However, finally extended at 72°C for 10 min. The PCR products were when the reaction proceeded for about 13 d, the complete then recovered using an AxyPrep DNA Gel Recovery Kit yellow-brown biofilm was observed on the surface of the 1# (Axygen), eluted with Tris-HCl, and detected by 2% agarose and 2# packings, which played a major role in the removal of NH+–N. Therefore, the removal rates of NH+–N began to gel electrophoresis. Based on the results of electrophoresis, 4 4 the PCR products were quantified by QuantiFluor TM-ST gradually increase. Studies have shown that the removal of (Promega (Beijing) Biotechnology Co., Ltd., No. 907-909, ammonia nitrogen at the beginning of the experiment relies Block B, Global Trade Center, 36 North Third Ring Road East, mainly on packing adsorption and physical retention [20]. Dongcheng District, Beijing), and the corresponding pro- The surface of the packing in the early stage of the reaction portions were mixed according to the sequencing amount of is rough, having many voids, large specific surface area, each sample [15]. The Illumina platform library was con- irregular particles, and can fully contact and react with the structed after fluorescence quantification. organic matter in the sewage to increase the retention of pollutants [21]. + 3. Results and discussion The NH4–N effluent concentration of the two packings generally decreased gradually, but only that of the 2# pack- 3.1. Influence of packing type on treatment effect ing could reach and stabilize below 5 mg/L, which satisfied the first-grade A discharge standard. This indicates that the 3.1.1. Influence of NH+–N removal effect 4 2# packing is feasible to treat low-concentration ammonia-ni- + The effect of the two packings on NH4–N removal is trogen wastewater. The reason may be that low DO inhib- + shown in Fig. 4. The removal rate of NH4–N fluctuated its the activity and proliferation rate of nitrifying bacteria. 192 C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196

2# packing can form the ideal state of coexistence of various expanding the range of anaerobic layers in the biofilm micro- microorganisms due to its reticulated three-dimensional environment of the 2# packing. A large number of heterotro- structure. Its surface and center can also be biologically phic nitrifying bacteria suitable for growth under low DO denitrified by aerobic microorganisms and anaerobic micro- were grown on the surface of the packing, and the anoxic organisms, respectively [22], which is more suitable for the microenvironment formed under low DO conditions makes reproduction of nitrifying bacteria. the denitrification process easier [24]. On the other hand, in the biological contact oxidation reactor, the anoxic tank was 3.1.2. Influence of TN removal effect not disposed, so the removal of TN mainly depended on the simultaneous nitrification and denitrification reaction occur- The effect of the two packings on TN removal is shown in ring on the biofilm. Ammonia oxidizing bacteria have higher Fig. 5. The removal rate of TN of the two packings generally affinity for oxygen than nitrite oxidizing bacteria, so low shows a trend of decreasing first, then increasing and finally DO conditions are beneficial to the short-range nitrification stabilizing. Moreover, the removal rate of TN of 2# packing reaction [25]. was higher than that of 1# packing, and the results are similar to those of Fig. 4. The lowest removal rate of 1# packing was 3.1.3. Influence of COD removal effect 47.92%; the highest removal rate was 55.32%, and the final removal rate was stabilized at 54.74%. The lowest removal The effect of the two packings on COD removal is shown rate of 2# packing was 48.85%; the highest removal rate was in Fig. 6. Both packings have high COD removal rates. The 63.51%, and the final removal rate was stabilized at 61.21%. lowest removal rate of 1# packing was 80.41%; the highest The TN removal rates of the two packings were not high, removal rate was 89.14%, and the final removal rate was sta- which is similar to the related research results. Wang et al. bilized at 87.41%. The lowest removal rate of 2# packing was [23] found that when the DO is less than 2 mg/L, the TN 78.32%; the highest removal rate was 92.14%, and the final removal rate of the biological contact oxidation method is removal rate was stabilized at 91.23%. low. The reason may be that even under low DO conditions, The DO concentration limits the microbial activity and the anoxic environment required for denitrification cannot be various biochemical reactions, and therefore has an import- satisfied. It is difficult for oxygen molecules to replace nitrate ant effect on the removal of organic matter. Both sets of reac- nitrogen as an electron acceptor, thereby inhibiting denitrifi- tors had a high removal rate of COD. On the one hand, under cation of denitrifying bacteria. In addition, the lack of organic the condition of low DO, the area of the anoxic layer in the carbon sources in low C/N domestic wastewater is also one membrane expands, which is more suitable for the growth of the reasons for the low TN removal rate [20]. of facultative heterotrophic bacteria. In the reactor system, It can be seen from Fig. 5 that the TN removal rate of the the removal of organic matter was originally mainly caused 2# packing was higher than that of the 1# packing. Only the by the decomposition of aerobic heterotrophic bacteria, and TN effluent concentration of the 2# packing reactor could then mainly by the facultative heterotrophic bacteria. This is be lower than 15 mg/L, which reached the first-grade A consistent with the findings of Wang et al. [26]. On the other discharge standard. On the one hand, because the porosity hand, COD removal relies mainly on the physical adsorption and specific surface area of the 2# packing were higher than and retention of the packing, as well as the synergistic effect those of the 1# packing, the 2# packing could enhance the of heterotrophic microbial growth and reproduction [27]. diffusion and mass transfer of DO in the biofilm under low In this study, the special mesh structure of 2# packing and oxygen conditions, thereby increasing the nitrification ability. the dense branch structure of 1# packing were beneficial to A larger specific surface area means more biofilm, thereby retain organic matter and increase the mass transfer rate of

35 101000 100100 140 30 95 80 120 90 25 100 60 85 20 80 80 InfluentInfluent watewaterr 15 TN(mg/L) TN(mg/L) 40 60 1#1# effluenteffluent water COD (mg/L) COD (mg/L) 2#2# effluenteffluent water 75 Removal rate(%) Removal rate(%) Removal rate(%) 1#1# reremoval rate Removal rate(%) 10 10 2#2# reremoval rate InfluentInfluent watewaterr 40 70 1#1# effluenteffluent watewater 20 2#2# effluenteffluent watewater 5 1#1# remremoval rate 20 65 1#1# remremoval rate 0 0 0 60 0102030405060 0102030405060 Day Day

Fig. 5. Effect of two kinds of packings on the removal of TN from Fig. 6. Effect of two kinds of packings on the removal of COD wastewater. from wastewater. C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196 193 oxygen. It could provide sufficient nutrients for the growth adopted by this research. (1) The adsorption and retention and attachment of microorganisms and meet the metabolic of fillers have certain effect on the removal of insoluble phos- activities of heterotrophic bacteria. phorus in sewage. (2) The metabolism of microorganisms needs to absorb phosphorus in the process of growth, that is, 3.1.4. Influence of TP removal effect assimilation. (3) Alkalinity will be produced in denitrification process. The precipitation effect of some phosphorus occurs The effect of the two packings on the removal of TP is in the process of pH change. Among them, microbial assimi- shown in Fig. 7. As can be seen from Fig. 7, the removal rates lation played the most important role in the removal process of TP by the two packings fluctuated greatly, showing a trend of this system [29]. of decreasing first and then stabilizing overall. The final Based on the results of Figs. 4–7, 2# packing was gener- + removal rate of TP of 1# and 2# packings was similar, and ally better than 1# packing for the removal of NH4–N, TN, the removal rate of TP of 2# packing was more stable. Among COD and TP, and the removal of pollutants from wastewater them, the lowest removal rate of 1# packing was 19.13%; the disposal after aeration. The main reason is that the surface highest removal rate was 53.17%, and the final removal rate of the 2# packing has some hydrophilic hydroxyl, cation and was about 27.96%. The lowest removal rate of 2# packing was other reactive groups, which can be combined with nega- 20.97%; the highest removal rate was 37.72%, and the final tively charged microorganisms to produce a stable chemical removal rate was about 27.89%. bond and valence combination. This allows the biofilm to The removal of phosphorus by biofilm process is usu- adhere firmly to the surface of the packing, which facilitates ally achieved by combined treatment process. Phosphorus the growth of the biofilm and enhances the removal of con- removal by the combined process of attachment growth and taminants. Table 4 shows the comparison between effluent suspension growth is usually chemical rather than biologi- water quality indexes of the wastewater treated by the two cal. This is because most biological phosphorus removal pro- packings and some water quality indexes of the first-class cesses require an anaerobic reactor at the front end. For most A emission standard of Discharge standard of pollutants combined processes, this requires the backflow of mixed for municipal wastewater treatment plant (GB 18918-2002). + liquor in the biofilm reactor, which is usually difficult to The removal efficiency of NH4–N, TN and COD in No. 2 achieve. Some researchers have also explored the phospho- reactor was better than that of No. 1 reactor, and only the rus removal effect of denitrifying phosphorus accumulating effluent indexes of No. 2 reactor could reach the discharge bacteria in the combined treatment process. For example, standard. The biological phosphorus removal effect of the in Zhang et al.’s [28] A2/O-BCO double sludge system, by single-contact oxidation process was not ideal and could be refluxing the nitrate from the contact oxidation stage into combined with anaerobic reactor. the anoxic tank, a large amount of NOX–N is provided as an electron acceptor for denitrifying phosphorus accumulating 3.2. Analytical results of microbial composition bacteria to achieve the phosphorus accumulating effect. The experimental device in this study was a single-con- In order to investigate the effects of different packings on tact oxidation process. The process did not have the condi- biomes in low C/N domestic sewage, the biofilm on the sur- tions of anaerobic/aerobic alternate operation, and there face of the stationary phase was sampled. The high-through- was no sludge system, so the effect of biological phospho- put sequencing of the surface biofilms of the two packings rus removal was not ideal. According to the analysis, there (labeled 1# packing and 2# packing) was performed using are three main ways to remove phosphorus in the system the MiSeq platform to obtain 36,512 and 35,789 optimized sequences, respectively. The results are shown in Table 5. The two sets of optimized sequences were intercepted and aligned 7 100 with the SILVA106 library, and then clustered. Under the condi- Infulent water 1# effluent water tion of 97% similarity, 927 and 1,035 operational taxonomic 90 6 2# effluent water units (OTUs) were obtained in the optimized sequences. The 1# removal rate 2# removal rate 80 Chao index indicates that the larger the richness index of the 5 microbial community in the sample, the greater the abun- 70 dance of the microbial flora. The Shannon index indicates

4 60 that the larger the diversity index of the microbial community in the sample, the more abundant the microbial diversity 3 50 contained in the sample. And the size of the coverage value TP (mg/L) TP (mg/L) reflects the proximity of the test results to the sample. Removal rate (%) 40 Removal rate (%) 2 It can be seen from Table 5 that the coverage rates of 30 the two groups of experimental samples were 99.44% and 99.49%, respectively, both greater than 95%, indicating that 1 20 the sequence library constructed in this study could cover the diversity of bacterial communities. In addition, the exper- 0 10 0102030405060 imental results of the two packings showed that: OTUs were 2# packing > 1# packing, Chao index was 2# packing > 1# pack- Day ing, Shannon index was 2# packing > 1# packing. Therefore, Fig. 7. Effect of two kinds of packings on the removal of TP from under the condition of low DO (DO = 1.5–2.0 mg/L), the 2# wastewater. packing had higher microbial richness and diversity index, 194 C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196

Table 4 Comparison between effluent water quality indexes and some water quality indexes in the first-class A emission standard of Dis- charge standard of pollutants for municipal wastewater treatment plant (GB 18918-2002)

Indexes 1# average 2# average First-class A emission standard of Discharge standard of pollutants (mg/L) (mg/L) for municipal wastewater treatment plant (GB 18918-2002)

+ NH4–N 5.15 3.11 5 TN 16.36 13.28 15 COD 16.99 11.88 50 TP 3.19 3.14 0.5

Table 5 100 Microbial richness and polymorphism index analysis table 90 other 80 Caldiserica Latescibacteria Sample 1 2 70 Firmicutes Coverage rate, % 99.44 99.49 60 Ignavibacteriae Gemmatimonadetes 50 OTUs 927 1,035 50 Verrucomicrobicrobiaa 40 Planctomycetes Chao index 999.11 1,089.87 Nitrospirae 30 Acidobacteria Relative abundance (%) Shannon index 7.27 8.36 Relative abundance (%) 20 Chloroflexi Bacteroidetes 10 ProteobacteriProteobacteriaa

0 which indicates that 2# packing is more suitable for the 1# packinpackingg22## packinpackingg growth and reproduction of microorganisms, while increas- + Fig. 8. Community structure diagram at the phylum level. ing the removal rate of COD, TN and NH4–N.

100 other 3.3. Analysis results of community structure Planctomycetacia 90 OM19OM1900 Caldisericia 80 Caldisericia Based on the classification information of SILVA data- PhycisphaeraPhycisphaeraee base, the high-throughput sequencing data of the attached 70 BacteroidiaBacteroidia Solibacteres colonies on the two kinds of packings were classified at the 60 Chlorobia Ignavibacteria phylum, the class and the genus level, and the community 50 50 Blastocatellia structure maps were analyzed which are shown in Figs. 8–10. 40 Gemmatimonadetes At the phylum category level shown in Fig. 8, there were Flavobacteriia 30 NitrospirNitrospiraa Relative abundance (%) 12 large groups of biofilm samples on the surface of the Relative abundance (%) Anaerolineae 20 Alphaproteobacteria two packings (relative abundance greater than 1%), includ- Sphingobacteriia 10 ing Proteobacteria (61.68%, 52.91%), Bacteroidetes (7.96%, Deltaproteobacteria 0 Gammaproteobacteria 11.43%), Chloroflexi (5.06%, 7.13%), Acidobacteria (5.44%, 1# packingpacking22## packinpackingg Betaproteobacteria 4.81%), Nitrospirae (3.46%, 5.55%) and Planctomycetes (3.12%, 4.51%). Fig. 9. Community structure diagram at the class level. The dominant phylum was Proteobacteria, which is the + largest species of bacteria in the phylum level and contains phenomenon that the effluent concentration of NH4–N and a wide variety of bacterial species, most of which are fac- TN of the 2# packing was lower than that of the 1# packing. ultative or obligate anaerobic heterotrophic bacteria. 67.15% At the class category level shown in Fig. 9, the sur- of Proteobacteria are Betaproteobacteria, and most of them face biofilm samples of the two packings belonged to 18 tend to acquire nutrients by decomposing organic mat- groups (relative abundance greater than 1%), including Beta ter under anaerobic conditions. Some Betaproteobacteria proteobacteria (33.08%, 24.72%), Gammaproteobacteria (7.81%, can utilize hydrogen, ammonia, methane and volatile fatty 5.50%), Deltaproteobacteria (13.89%, 15.85%), Sphingo acids, which are thought to be closely related to sludge bacteria (6.45%, 9.73%), Alphaproteobacteria (6.60%, 5.89%) denitrification [14,30]. and Anaerolineae (2.95%, 4.77%). The dominant class of micro- Under the low DO conditions of the experimental study, organisms on both packings were Betaproteobacteria, and the microorganisms on the packing were mainly faculta- some species of Betaproteobacteria are related to the denitri- tive heterotrophic bacteria. The heterotrophic bacteria had fication process [31,32]. Although the Betaproteobacteria a significant effect on organic pollutants removal in waste- richness of the 1# packing was greater than that of the 2# water, which is consistent with the results of the above two packing, the premise of denitrification is to have a good experiments. It is worth noting that the relative richness of nitrification reaction, and the experimental results of the 2# Nitrospirae of the 2# packing was greater than that of the packing (shown in Fig. 4) met this requirement. The nitrifica- 1# packing. Nitrospirae is the main nitrous acid oxidizing tion of 2# packing was better than that of 1# packing, and the – – bacteria, which oxidizes NO2 to NO3. This also explains the removal rate of TN was also higher. C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196 195

100 combined packing reactor, and only the effluent indexes 90 other of polyurethane suspension packing reactor could reach Flavobacterium 80 80 Phaeodactylibacter the first-class A emission standard of Discharge standard 70 Ferruginibacter of pollutants for municipal wastewater treatment plant OM27_clade (GB 18918-2002). The removal of biological phosphorus 60 Sphaerotilus uncultureunculturedd of the single-contact oxidation process was not ideal and 50 Hydrogenophaga could be combined with anaerobic reactor. Under low 40 Desulfobacterium Nitrospira DO, the use of polyurethane suspension packing by bio- 30 Relative abundance (%) Relative abundance (%) Denitratisoma logical contact oxidation method was feasible in remov- 20 Thiothrix + unidentified ing domestic sewage NH4–N, TN and COD. 10 • The richness and diversity of bacterial communities in 0 1# packinpackingg22## packing polyurethane suspension packing reactor were higher than those of composite packing reactor. Fig. 10. Community structure diagram at the genus level. • The dominant flora in the two groups was Proteobacteria, Betaproteobacteria and Denitratisoma. Polyurethane sus- At the genus category level shown in Fig. 10, there were pension packing reactor contained relatively richer nitri- 10 large groups (relative abundance greater than 1%) in the fication-related flora Nitrospira, which provide better biofilm samples attached to the two packings, including growth and reproduction conditions for the denitrifica- Thiothrix (2.52%, 0.54%), Denitratisoma (4.20%, 9.20%), Nitrospira tion-related flora Betaproteobacteria and Denitratisoma. (3.44%, 5.51%) and Desulfobacterium (1.25%, 2.79%). The dom- The large amount of facultative heterotrophic bacteria in inant genus of both experimental packings was Denitratisoma, the Proteobacteria had a significant effect on the removal which is similar to the related research results. Zhao et al. of organic pollutants. [14] found Denitratisoma (3.71%) associated with denitrification in the study of denitrification and microbial community Acknowledgment characteristics of composite carbon source packings. The results of Fahrbach et al. [33] showed that This work was supported by Marine public welfare indus- Denitratisoma isolated from activated sludge from municipal try research special fund projects (201405035 and 201505006). sewage treatment plants is a newly discovered genus belong- ing to Rhodocyclaceae. Denitratisoma has the ability of aer- References obic denitrification and short-range nitrification, which can – [1] F. Ding, L.Y. Huang, R. Wang, Y.J. Gao, J.Y. Yao, X.H. Wang, Q. Li, directly convert NO2 into gaseous nitrogen. The relative richness of Denitratisoma in 2# packing was greater than that in 1# Water pollution emergencies in China, 2004–2015: monitoring data analysis, Chin. J. Public Health, 33 (2017) 59–62. packing, which indicates that 2# packing can provide a good [2] F.Y. Chen, H.W. Zhou, Safety situation and protection counter environment for microbial growth, strengthen the bioenrich- measures of drinking water sources in Taihu Basin, Water Dev. ment of Denitratisoma and make TN removal better than that Res., 15 (2015) 16–18. of 1# packing (shown in Fig. 5). In addition, Nitrospira is a [3] R.R. Hou, S.J. Guo, H.L. Peng, Y.F. Chen, Effects of DO and reflux genus of nitrifying bacteria, and the contents of Nitrospira in ratio on removal of nitrogen and phosphorus in integrated sewage treatment plant, Modern Chem. Ind., 38 (2018) 87–91. the #1 and 2# packings were 3.44% and 5.51%, respectively. [4] Y.Z. Peng, B. Bin, Review of biological nitrogen removal This also creates the possibility of Nitrospira’s enrichment. enhancement technologies and processes under low C/N ratio, Liu and Wang [34] also confirmed that long-term operation Acta Sci. Circumst., 29 (2009) 225–230. under low DO conditions can enrich Nitrospira [35]. [5] Q. Zhang, S.Y. Wang, Y.Y. Miao, X.X. Wang, Y.Z. Peng, Start-up As can be seen from Figs. 8–10, the dominant flora in the of CANON process on domestic wastewater using intermittent aeration with low DO, CIESC J., 68 (2017) 289–296. two groups of reactors was Proteobacteria, Betaproteobac [6] X.L. Song, X. Zhang, J.R. Li, D. Niu, Study on treatment of teria and Denitratisoma. Polyurethane suspension pack- domestic sewage by biological contact oxidation process, Chem. ing reactor contained relatively richer nitrification-related Bioeng., 24 (2007) 66–68. flora Nitrospira, which provide better growth conditions for [7] M. Di, C.J. Xu, S.J. Zhao, F. Wang, H.Y. Peng, W.Q. Huan, Air the denitrification-related flora Betaproteobacteria and Floating/Hydrolysis Acidification/Biological Contact Oxidation/ Filtration Process for Treatment of ASP Flooding Produced Denitratisoma. This is consistent with the better removal Water, China Water & Wastewater, 32 (2016) 69–70. + of NH4–N and TN in polyurethane suspension reactor [8] H.G. Yin, Z.X. Tang, X. Wang, Y.M. Wang, Dyeing wastewater (Figs. 4 and 5). The large amount of facultative heterotro- purification by using ABR and two steps biological contact phic bacteria in Proteobacteria had a significant effect on oxidation tank, Dyeing Finishing, 3 (2017) 268–272. organic pollutants removal, which is consistent with the [9] A. Pallavi, K. Sunil, K. Kavita, Studies on packed bed treatment for organic matter in distillery effluent, Int. J. Eng. Sci. Innov. COD removal result in Fig. 6. In the future, Proteobacteria, Technol., 3 (2014) 268–272. Betaproteobacteria, Denitratisoma and Nitrospira flora can be [10] M.A. El-Khateeb, M.A. Saad, H.I. Abdel-Shafy, F.A. Samhan, cultivated in a targeted manner to improve the treatment of M.F. Shaaban, The feasibility of using non-woven fabric as low C/N domestic sewage under low DO conditions. packing material for wastewater treatment, Desal. Wat. Treat., 111 (2018) 94–100. [11] Y. Tanji, R. Sakai, K. Miyanaga, H, Unno, Estimation of the 4. Conclusion self-purification capacity of biofilm formed in domestic sewer pipes, Biochem. Eng. J., 31 (2006) 96–101. • The polyurethane suspension packing reactor had bet- [12] J.C. Zhang, X.D. Hu, Y.F. Shi, Z.M. Sun, Y.K. Zhang, Comparative + study of different packing biological contact oxidation processes ter removal efficacy on NH4–N, TN and COD than the 196 C.H. Shi et al. / Desalination and Water Treatment 181 (2020) 188–196

for organic wastewater treatment, Water Wastewater Eng., [24] J.L. Wang, S.Y. Wang, Y.Z. Peng, Y.Q. Gao, Z.R. Sun, Simultaneous 3 (2015) 136–139. nitrogen and phosphorous removal under low DO in a hybrid [13] Z.Q. Shen, Y.X. Zhou, J. Hu, J.L. Wang, Denitrification per biological reactor, Environ. Sci., 29 (2008) 655–659. formance and microbial diversity in a packed-bed bioreactor [25] X.W. Zhou, Y.M. Lu, L. Chen, Q. Wang, W.L. Jia, Effect of using biodegradable polymer as carbon source and biofilm magnetic field on running performance of shortcut nitrification support, J. Hazard. Mater., 250–251 (2013) 431–438. process, Environ. Eng., 35 (2017) 34–37. [14] W.L. Zhao, R.X. Hao, R.Z. Wang, P. Du, Denitrification of [26] L. Wang, H.Y. Li, X.J. Li, F. Ke, J.Z. Peng, Study on the effects of composite carbon filler and character of microbial community, low dissolved oxygen on biofilm characteristics, China Water China Environ. Sci., 35 (2015) 3003–3009. Wastewater, 24 (2018) 15–19. [15] F.F. Li, C.H. Shi, H.B. Li, R.F. Yuan, F.S. Ma, Impact of [27] Y. Ge, Studies on the Function of Substrates and pollutants starvation conditions on biological community structure in removal mechanisms in Subsurface Flow Constructed Wetlands, sulfur autotrophic denitrification reactor, Environ. Sci., 38 (2017) Xi’an University of Architecture and Technology, Xi’an, 2017. 1109–1115. [28] M. Zhang, Q. Yang, J.H. Zhang, C. Wang, S.Y. Wang, Y.Z. Peng, [16] L. Cai, Research on the Application Characteristics of Enhancement of denitrifying phosphorus removal and microbial Different Fillers in Biological Contact Oxidation and Process community of long-term operation in an anaerobic anoxic Improvement, Harbin Institute of Technology, Harbin, 2011. oxic–biological contact oxidation system, J. Biosci. Bioeng., 122 [17] Y. Du, B.H. Zhou, R.F. Yuan, X.M. Bao, D.W. Li, Treatment of (2016) 456–466. large-scale livestock wastewater by UASB-SBR process, Chin. J. [29] G.L. Qiu, Study of Biological Aerated Filter Process and Environ. Eng., (2018). Biological Contact Oxidation Process Used in Decentralized [18] M.R. Gillings, Rapid extraction of PCR-competent DNA from Wastewater Treatment, Shandong University of Science and recalcitrant environmental samples, Methods Mol. Biol., 1096 Technology, Qing Dao, 2007. (2014) 17. [30] T.R. Thomsen, Y.H. Kong, P.H. Nielsen, Ecophysiology of [19] Zielińska, Sylwia, P. Radkowski, A. Blendowska, A. Ludwig- abundant denitrifying bacteria in activated sludge, Fems Micro Gałęzowska, The choice of the DNA extraction method may biol. Ecol., 60 (2010) 370–382. influence the outcome of the soil microbial community structure [31] Y. Chen, B.S. Sun, B. Zhang, X. Huang, Nitrifying bacteria analysis, MicrobiologyOpen, 453 (2017). structure community of different MBR reactor, China Environ. [20] L.H. Kong, C.H. Shi, F.S. Ma, B.H. Zhou, Comparison of Sci., 30 (2010) 69–75. treatment on SBR tail water in tidal flow constructed wetlands [32] F. Huang, X.J. Mei, Z.W. Wang, Z.C. Wu, Diversity of operation with different packing, Chinese J. Environ. Eng., 11 (2017) performance and microbial community structures in MBR and 379–385. CAS processes at low temperature, Environ. Sci., 35 (2014) [21] L.T. Pan, X.J. Xie, H.C. Wang, H.J. Dong, Preparation of 1002–1008. biological filter packing for nitrogen and phosphorus removal [33] M. Fahrbach, J. Kuever, R. Meinke, P. Kampfer, J. Hollender, and its treatment effect on rural domestic sewage, Trans. Denitratisoma oestradiolicum gen. nov., sp. nov., a 17beta- Chinese Soc. Agric. Eng., 33 (2017) 230–236. oestradiol-degrading, denitrifying betaproteobacterium, Int. J. [22] G.Y. Sun, E.F. An, Y. Liu, H.F. Li, Study on membrane-forming Syst. Evol. Microbiol., 56 (2006) 1547–1552. and treatment effect of different biofilm carriers for micro- [34] G.Q. Liu, J.M. Wang, Long-term low DO enriches and shifts polluted water purification, Water Wastewater Eng., 43 (2017) nitrifier community in activated sludge, Environ. Sci. Technol., 48–53. 47 (2013) 5109–5117. [23] Y.C. Wang, W. Chen, X.Y. Zheng, W.L. Huang, J. Sun, Influencing [35] Y.Y. Gao, Y.Z. Peng, P. Bao, S.Y. Guo, S.Y. Wang, The characteristic factors of simultaneous nitrification and denitrification in of activated sludge in nitrifying low-DO reactor, China Environ. biological contact oxidation process, China Water Wastewater, Sci., 37 (2017) 1769–1774. 27 (2011) 22–25.